Upgrade to Pro — share decks privately, control downloads, hide ads and more …

American University Colloquia

American University Colloquia

We have seen the discovery and confirmation of thousands of exoplanets since the first planet found orbiting a Sun-like star, and we are now on the verge of entering an exciting new era of planetary exploration: detection and characterization of terrestrial exoplanet atmospheres. Detecting H2O, O2, and O3 in exoplanet atmospheres is the first step on the path to determining planet habitability, and efficiency is key to maximizing the science output from limited observation time, especially in next-generation instrument design such as the upcoming Habitable Worlds Observatory (HWO). Knowing this, the optimal wavelength for the spectral bandpass used for observations is a crucial factor to consider.

Coronagraphic design currently limits the observing strategy used to detect H2O, O2, and O3, requiring the choice of specific bandpasses to optimize abundance constraints. We use a pre-constructed grid consisting of 1.4 million geometric albedo spectra across a range of abundance and pressure and interpolate to produce forward models for an efficient nested sampling routine, PSGnest, thus enabling wide ranges of parametric retrievals. By understanding the SNR requirements for detecting molecules of interest, and properly prioritizing the spectral bandpasses to optimize detectability of different atmospheric constituents, we can inform the best instrument designs and observing procedure as we look to the HWO.

HWO is a multi-generational Great Observatory telescope, encompassing many years of scientists being trained and recruited. By implementing effective and ethical mentorship techniques based on current research into the ground floor of HWO development, we can increase the rate of recruitment and retention of historically minoritized groups in astronomy and physics using HWO.

Natasha Latouf

October 04, 2023
Tweet

Other Decks in Science

Transcript

  1. Natasha Latouf
    George Mason University | September 20th, 2023 | American University
    Using Grid-Based Nested Sampling in
    Coronagraphy Observation Simulations for
    Molecular Detection

    View Slide

  2. Introduction
    ★ Since the first exoplanet was found almost 30 years ago,
    we have discovered and confirmed over 5,000
    exoplanets orbiting host stars.
    ★ Detecting and characterizing terrestrial exoplanet
    atmospheres is the next step to learning about Earth’s
    development.
    ★ With proposed future missions, we can probe for
    biosignatures that can hint at the likelihood of clement
    conditions and biological activity (Kaltenegger 2017,
    Schwieterman et al. 2018).

    View Slide

  3. Okay, so why do I care now?
    ★ Coronagraphy allows us to directly image a
    planet’s light.
    ★ The inner working angle (IWA) of coronagraphs is
    proportional to wavelength, and favors shorter
    wavelengths. However, some biosignatures have
    prominent features at longer wavelengths.
    ★ By understanding the requirements needed for
    detection, we can have efficient observing
    procedures which can inform telescope design
    and extend the lifetime of the observatory. By Jason Wang (Caltech)/Christian Marois (NRC
    Herzberg), CC BY 4.0

    View Slide

  4. BARBIE 1: Motivation
    ★ Detecting H2
    O stands as the first
    step in the search for habitability
    ★ Multiple factors affect the
    efficiency of atmospheric
    characterization at varying
    wavelengths, including SNR
    ★ Varying the molecular abundance
    allows us to study multiple Earth
    epochs

    View Slide

  5. BARBIE 1: Methodology
    Inputs:
    ★ Pre-computed spectral grid from
    Susemiehl et al. 2023 consisting of 1.4
    million geometric albedo spectra.
    ★ Fiducial Spectrum, split into 25 bandpasses
    from 0.515 - 1 µm
    ○ All true values set to modern Earth in first
    tests
    ○ For abundance case studies, all modern
    Earth values except H2
    O
    Input to nested sampling routine PSGnest using
    GridRunner

    View Slide

  6. BARBIE 1: Methodology
    Nested Sampling:
    1. Place a number of “live points” throughout a
    (given) prior space, and a likelihood is
    calculated
    2. Live point with the lowest log-likelihood is
    discarded
    3. A new, unrelated live point is placed in the now
    restricted prior space
    4. Repeat until remaining prior space is negligibly
    small
    PSGnest:
    ★ Housed in the Planetary Spectrum Generator
    (PSG)
    ★ Based on MultiNest

    View Slide

  7. BARBIE 1: Methodology
    Outputs:
    ★ Highest-likelihood values of output
    parameters,
    ★ Posterior distribution,
    ★ Uncertainties (calculated from posterior
    distribution),
    ★ Bayesian log-evidence (logZ)
    The Bayes Factor (lnβ):
    ★ Calculated by subtracting the Bayesian
    log-evidence per retrieval.
    ★ The comparison is between a retrieval
    with all parameters included vs a retrieval
    with one parameter removed.

    View Slide

  8. BARBIE 1: Modern Earth Results
    Now zooming in:

    View Slide

  9. BARBIE 1: Modern Earth Results
    Minimum H2
    O Bandpasses:
    ★ H2
    O is strongly detectable at all SNR ≥ 5
    and weakly detectable at all SNR at 0.9
    µm
    ★ H2
    O is strongly detectable at 0.74 µm at
    SNR ≥ 12
    Multi-Species Constraints:
    ★ Included retrieval for Modern Earth
    abundance of O2
    ★ Results suggest that an SNR = 10
    spectrum at 0.83 µm could achieve a
    strong detection of both H2
    O and O2

    View Slide

  10. BARBIE 1: Abundance Case Studies
    Remember: lnβ is detectability strength!

    View Slide

  11. BARBIE 1: Abundance Case Studies
    Wavelength Detectability:
    ★ H2
    O is only detectable at 0.9 µm and
    0.85 µm for abundances < 1.7e-3 VMR,
    and only at high SNR
    ★ When H2
    O ≥ 1.7e-3 VMR, it is detectable at
    all wavelengths of interest, with varying
    SNR
    Observing Earth’s Epochs:
    ★ Neoproterozoic unlikely to be detected
    ★ Mid-Cretaceous Greenhouse and
    Modern more likely to be detected
    especially at higher abundances

    View Slide

  12. BARBIE 1: Published!

    View Slide

  13. BARBIE 2: Motivation
    ★ O2
    and O3
    have geochemical
    proxies that allow for better
    estimation of abundances through
    time
    ★ Studying the oxygenation of
    Earth’s surface can inform the
    global biochemical development

    View Slide

  14. BARBIE 2: Modern Earth Results, 20% BP

    View Slide

  15. BARBIE 2: Modern Earth Results, 20% BP
    Minimum O2
    , O3
    Bandpasses:
    ★ O2
    is strongly detectable at 0.74 µm at
    SNR ≥ 8
    ★ O3
    is strongly detectable at 0.63 µm at
    SNR ≥ 19, and weakly detectable at 0.60
    µm at SNR ≥ 14
    Multi-Species Constraints:
    ★ An SNR = 10 spectrum at 0.83 µm could
    achieve a strong detection of H2
    O & O2
    ★ An SNR = 17 spectrum at 0.65 µm could
    achieve a strong detection of H2
    O and a
    weak detection of O3

    View Slide

  16. BARBIE 2: Abundance Case Studies, 20% BP
    ★ O2
    is not detectable lower than 50% PAL O2
    , therefore it is unlikely that a Proterozoic
    abundance level of O2
    can be detected.
    ★ O3
    is detectable across the entire range of values, with moderate-high SNR data for
    Modern abundance levels and low SNR data for Early abundance levels.
    Proterozoic

    View Slide

  17. BARBIE 2: Bandpass Comparison
    ★ O2
    and O3
    have geochemical
    proxies that allow for better
    estimation of abundances through
    time
    ★ Studying the oxygenation of
    Earth’s surface can inform the
    global biochemical development
    ★ Varying the bandpass widths
    allows us to study the change in
    detectability strength.

    View Slide

  18. BARBIE 2: Bandpass Comparison
    ★ O2
    detectability does not change significantly across bandpass width.
    ★ O3
    detectability changes significantly across bandpass width.
    ★ Dual or triple detections are now possible!

    View Slide

  19. BARBIE 2: Abundance Case Studies

    View Slide

  20. BARBIE 2: Thermochemical Consistency
    O2
    /O3
    Coupled Abundance:
    ★ By shifting the abundance of one
    molecule at a time, it is not chemically
    accurate.
    ★ Using the Kozakis et al. 2022 Figure 2,
    we paired O2
    /O3
    abundances to mimic a
    more chemically accurate atmosphere.
    Abundance Pairings:
    ★ 100% O2
    PAL/100% O3
    PAL
    ★ 50% O2
    PAL/110% O3
    PAL
    ★ 10% O2
    PAL/105% O3
    PAL Kozakis et al. 2022, Figure 2

    View Slide

  21. BARBIE 2: ATMOS Convergence

    View Slide

  22. BARBIE 2: ATMOS Convergence

    View Slide

  23. BARBIE 2: ATMOS Convergence
    ★ H2
    O limits dual or triple detection with O2
    in both SNR and wavelength
    ★ H2
    O limits dual or triple detection with O3
    in SNR, but O3
    is limiting in wavelength
    ★ Triple molecular detection is possible at 30% and 40% bandpasses, but at 40% there
    is little margin for single molecule detection.
    50% O2
    PAL/110% O3
    PAL

    View Slide

  24. Mentorship Motivation
    ★ The Habitable Worlds Observatory (HWO) will be a multi-decade effort
    spanning generations of scientists.
    ★ Historical and systemic barriers have resulted in extremely low rates of
    recruitment and retention for historically minoritized groups (HMGs) in
    physics and astronomy.
    ★ Mentorship was designated as a priority moving forward in the 2020
    Astrophysics Decadal Survey, with emphasis on research based techniques
    (Appendix 7).

    View Slide

  25. Mentorship Motivation
    Effective and ethical mentorship can:
    ★ Help students find a sustainable balance between their scientific research
    and their commitment to the field (Pando 2022)
    ★ Provide a sense of community and ownership (Godwin 2016)
    ★ Increase satisfaction with their career in the long term (Pando 2022)
    ★ Increase retention and prepare both mentors and mentees for their chosen
    careers in physics in the long term (Godwin 2016)

    View Slide

  26. Mentorship Future Works
    In the future, we will:
    ★ Present a meta-analysis of available mentorship research in the fields of
    physics and astronomy as pertaining to HMGs.
    ★ Provide a research-based, effective mentorship training and guidance
    development at the ground floor of any active mission concept.
    ★ Promote open communication and pairing through natural and planned
    mentorship.

    View Slide

  27. Conclusions
    ★ H2
    O, O2
    , and O3
    can be detected simultaneously as low as 0.66 µm at a 30%
    bandpass with moderate-high (≥13) SNR data.
    ★ By understanding the SNR and bandpass width requirements for detecting
    molecules of interest and properly prioritizing spectral bandpasses for
    optimal detectability, we can inform the best instrument designs and
    observing procedure.
    ★ The upcoming HWO presents an opportunity to introduce ethical and
    effective mentorship at the ground floor of development.

    View Slide

  28. Thank you!
    Science Collaborators:
    Dr. Avi Mandell, Dr. Gerónimo Villanueva,
    Michael Moore, Nick Susemiehl, Dr. Vincent
    Kofman, Dr. Michael Himes, Dr. Giada Arney,
    Jaime Crouse, Dr. Amber Young
    Mentorship Collaborators:
    Dr. Paula Danquah-Brobby, Dr. Jenna Cann
    Contact me:
    Email: nlatouf @ gmu.edu Twitter: @nertushka

    View Slide